Assessment

make this a question! How do Emergency Contraceptive Pill Prescribing Patterns very in Scotland?

Mark scheme: Justified – understanding of the data and their context
Introduction
Key results
Conclusion References (for reason chosen research question)
Narrative shows critical thinking
Multiple data sources
Recommendations and conclusions are grounded in the data
Limitations of the data set discussed. Next steps suggested in terms of data that would allow for further analysis.

Introduction

There are two types of emergency contraceptive pill prescribed by pharmacies.

• Levonogestrel (brand name: Levonelle) taken within three days of unprotected sex
• Ulipristal Acetate (brand name: ellaOne) taken within five days of unprotected sex

This report explores

1. Prescribing patterns of emergency contraceptive pill (ECP) between 2019 to 2023.

2. How ECP and antibiotics for STI prescriptions differ in geography comparing university town to a non-university town

3. Compare patterns in ECP and STI antibiotics prescribing during Covid-19 lockdowns - did the need for ECP or incidence of STI antibiotics prescription go down during lockdowns?

4. How does SIMD relate to prescription of ECP / antibiotics for STIs

# load necessary libraries
library(tidyverse)
library(janitor) 
library(gt) 
library(here)
library(lubridate)
library(patchwork)
library(plotly)

I have chosen to look at all prescriptions from 2020 and 2023. To do this I downloaded the prescriptions data for each month from: https://www.opendata.nhs.scot/dataset/prescriptions-in-the-community

I wanted to explore how many ECP and antibiotics for STIs were prescibed during across 2023.

4 plots and / or tables No use of settings other than default Labelled, titled plots Labelled titled tables Appropriate and thoughtful data visualisation with some use of non-default settings (gt) Faceting, multilayered plots, interactive visualisation

Read in data:

# Read in the Health Board names (HB_names). See code for link
HB_names <- read_csv("https://www.opendata.nhs.scot/dataset/9f942fdb-e59e-44f5-b534-d6e17229cc7b/resource/652ff726-e676-4a20-abda-435b98dd7bdc/download/hb14_hb19.csv") %>% 
  clean_names() # ensures column names are unique, lower case and replaces spaces and special characters with underscores. I will use this function for all read ins to ensure my data is consistent and easy to manipulate.


# Read in population data per Health Board from the 2022 census data. Available from: https://statistics.ukdataservice.ac.uk/dataset/scotland-s-census-2022-uv102a-age-by-sex/resource/b2d295c2-af53-4b3d-a075-7815cadd9060 
all_population_data <- read_csv(here("data", "UV103_age_health_board_census.csv"), skip = 10) %>% # locates the csv and excludes the first 10 lines in the csv as they are redundant
  rename(Spare = "...6", # remove unused columns
         hb_name = "Health Board Area 2019",
         hb_population = Count) %>% # rename() formats the data to match the prescriptions dataframe
  filter(Sex == "Female") %>% # filter female population, as men do not take ECP
  select(hb_name, Age, hb_population) %>% # select columns of interest
  mutate(hb_name = paste("NHS", hb_name)) %>%  # change hb_name column format to match the Health Board dataframe format
  clean_names()

# create dataset with total population for each health board
population_data <- all_population_data %>% 
  filter(age == "All people") %>% 
  select(hb_name, hb_population)

# select the population aged 16 to 34 years as this is the population the population most likely engaging with risk taking behaviors such as unprotected sex. I chose 16 as this is the age of consent in Scotland, and 34 as my cut off for young adults. 
age_population_data <- all_population_data %>% 
  filter(age%in% c(16:34)) %>% # filter ages of interest
  mutate(age = as.numeric(age)) %>% # make numerical so they can be placed in to buckets
  mutate(age_group = case_when(between(age, 16,34)~'16-34')) %>% # make bucket 16-34 years
  group_by(hb_name, age_group) %>% 
  summarise(pop_hb_16_34 = sum(hb_population)) # sum total population aged 16 to 34 per health board region. 

Define a function to read in the prescription data for a defined year:

# #For efficiency I created a function to read in my prescriptions datasets.
# #I downloaded 12 months of prescription data for each year from 2019 to 2022. I placed the 12 months of data into their relevant folder named all_months_year, where year was specific to the data it contained.
# read_all_prescriptions <- function(year){
#   all_files <- list.files(here("data", paste0("all_months_", year)), pattern = "csv") #list.files() retrieves files from the relevant directory, and the paste0() dynamically constructs the folder name based on the year variable placed into the function.
#   all_prescriptions <- all_files %>%
#     map_dfr(~read_csv(here("data", paste0("all_months_", year),.))) %>% #map_dfr() row-binds the datasets being red in
#     clean_names() %>%
#     drop_na(bnf_item_description) # drop the rows with missing bnf_item values
#   return(all_prescriptions)
# }

Read in datasets using read_all_prescriptions() function and start tidying:

# all_prescriptions_2019 <- read_all_prescriptions(2019)
# all_prescriptions_2020 <- read_all_prescriptions(2020)
# all_prescriptions_2021 <- read_all_prescriptions(2021)
# all_prescriptions_2022 <- read_all_prescriptions(2022)
# 
# # 2019 dataset has hbt2014 as a column name instead of hbt. Rename to make column name consistent
# all_prescriptions_2019 <- all_prescriptions_2019 %>%
#   rename(hbt = "hbt2014")
# 
# # combine 4 years of data into one dataset to make it easier to wrangle
# combined_prescriptions <- bind_rows(
#   all_prescriptions_2019,
#   all_prescriptions_2020,
#   all_prescriptions_2021,
#   all_prescriptions_2022 )%>% # I mutate early and select the prescriptions of interest to prevent a very large dataset from being stored in my environment (prevents R slowing down and crashing)
#   mutate(
#     date = parse_date_time(paid_date_month, "ym"), # use lubridate to format date
#     drug_simple = case_when(
#       str_detect(bnf_item_description, "LEVONO") ~ "Levonorgestrel",
#       str_detect(bnf_item_description, "ULIPR") ~ "Ulipristal Acetate",
#       TRUE ~ "Other")) %>% # used case_when() to group different dosages of the same drug
#   filter(drug_simple != "Other",!is.na(date)) %>% #remove prescriptions of no interest and any missing date values
#   filter(hbt != "SB0806") %>% # filter out SB0806 as it is not a health board (Scottish Ambulance Service)
#   filter(!is.na(hbt))

# # save combined_prescriptions dataset to csv
# # write_csv(combined_prescriptions,"data/combined_prescriptions.csv")

# filter drugs of intereset before completing read in to prevent loading large datasets into the environment (prevents R from running slow and crashing)
combined_prescriptions <- read_csv(here("data","combined_prescriptions.csv")) %>%
  mutate(
    date = parse_date_time(paid_date_month, "ym"), # use lubridate to format date
    drug_simple = case_when(
      str_detect(bnf_item_description, "LEVONO") ~ "Levonorgestrel",
      str_detect(bnf_item_description, "ULIPR") ~ "Ulipristal Acetate",
      TRUE ~ "Other")) %>% # used case_when() to group different dosages of the same drug
  filter(drug_simple != "Other",!is.na(date))%>%  # remove other prescriptions from the dataset and any missing date values
  filter(hbt != "SB0806") %>% # filter out SB0806 as it is not a health board (Scottish Ambulance Service)
  filter(!is.na(hbt))

Join and wrangle data:

# join the Prescriptions dataset to the Health Board names dataset and health board population data
ECP_scripts <- combined_prescriptions %>% 
  full_join(HB_names, by = c("hbt" = "hb")) %>% # Join with Health Board names
  full_join(population_data, by = "hb_name") %>% # Join with population data
  select(gp_practice, date, drug = drug_simple, hb_name, paid_quantity,hb_population) %>%  # select columns of interest
  mutate(month = factor(month(date), levels = 1:12, labels = month.abb), .after = date) %>%   # extract month as a factor with labels to help when plotting
  mutate(year = factor(year(date)), .after = month) 

Key Results

What are the seasonal trends in prescribing of Emergency Contracpetion Pills in the years 2019 to 2022?

This graph explores seasonal trends in prescribing of the two most commonly prescribed emergency contraceptive pills, Levonogestrel and Ulipristal Acetate, between 2019 to 2022 by Health Board. I plotted this to see if events such as Valentines day, University term start dates, or national holidays effect emergency contraceptive prescribing rates. I included 4 years of consecutive data to see if a one-time event, such as the Common Wealth Games caused any spikes in emergency contraception prescribing. Additionally by plotting the 4 consecutive years we can see if Covid-19 lockdowns in the years 2020 and 2021 affected emergency contraception prescribing rates.

Plot: Population-Weighted Rates: Ensures that national trends are calculated based on the total population across all health boards, avoiding bias from smaller regions. Accurate National-Level Insights: Better reflects overall prescribing behaviour at the national level, aligning with your goal to analyse seasonal and national events. prescribing rates are weighted by the total population, providing an accurate reflection of overall prescribing behaviour across Scotland.

total_population <- sum(population_data$hb_population, na.rm=TRUE)

ECP_presc_sum_month <- ECP_scripts %>%
  group_by(date, drug, month, year) %>%
  summarise(total_quantity_month = sum(paid_quantity, na.rm = TRUE)) %>% 
  ungroup() %>% 
  drop_na(drug) %>% 
  clean_names()

ECP_plot_data_v2 <- ECP_presc_sum_month %>%
  group_by(drug, year, month) %>% 
  summarise(prescriptions_per_100000 = (total_quantity_month / total_population)*100000) #%>%   #### 100,000
# labels / levels ? mutate( year = factor list c() ######### 

ECP_plot <- ECP_plot_data_v2 %>%
  ggplot(aes(x = month, y = prescriptions_per_100000, group = year)) +
  # Layer for other years
  geom_line(
    data = . %>% filter(year != 2020),
    aes(linetype = as.factor(year)),
    color = "grey70",  # Grey for other years
    size = 0.7
  ) +
  # Layer for 2020
  geom_line(
    data = . %>% filter(year == 2020),
    aes(linetype = as.factor(year)),
    color = "red",  # Highlight 2020 in red
    size = 0.9
  ) +
  facet_wrap(~drug, scales = "free_y") +
  scale_linetype_manual(
    values = c("2019" = "dotdash", "2020" = "solid", "2021" = "dashed", "2022" = "dotted"),
    name = "Year") +
  labs(
    title = ("Seasonal Trends in Emergency Contraception Prescribing (2019–2022): \n Accounting for Population-Adjusted National Prescribing Rates"),
    subtitle = "A four-year analysis of prescribing rates per 100,000 women, \n highlighting the impact of the Covid-19 lockdowns (March–July 2020, shown in red),\n with rates adjusted for population to ensure national-level accuracy",
    x = "Month",
    y = "Prescriptions per 100,000 women",
    #color = "Year Colour"
  ) +
  theme_minimal() +
  theme(
    plot.title = element_text(size = 12, hjust = 0.5,face = "bold"),
    plot.subtitle = element_text(size = 10, hjust= 0.5),
    axis.title.x = element_text(size = 10),
    axis.title.y = element_text(size = 10),
    axis.text.x = element_text(angle = 45, hjust = 1, size = 8),
    axis.text.y = element_text(size = 8),
    strip.text = element_text(size = 10),
    legend.title = element_text(size = 10),
    legend.text = element_text(size = 8),
    legend.position = "bottom"
  )

ECP_plot

investigate what happened in 2019 - change in prescription practises over time ? did they used to use it for something else?

Trends to note:

• Levonorgestrel is more commonly prescribed than Ulipristal Acetate.
• 2019 and 2022 represent years without Covid-19 lockdown measures. 2019 and 2022 have a peak in prescriptions of Levonorgestrel in October and August respectively. This perhaps reflects increased sexual activity coupled with the return to university with ‘fresher’ week.
• Between Febuary and May of 2020 the prescription rate of Levonorgestrel decreases. This is likely due to decreased sexual activity and risk taking behaviours due to the stringent national ‘lockdown’ measures introduced between X to Y in response to the Covid-19 pandemic.
• At the start of 2021 prescription rate of Levonorgestrel is very low. Likely due to additional Covid-19 lockdowns.

It would be interesting to explore if different Health Boards prescribe different amounts of ECP.

Which Health Board is prescribing the most ECP?

play around:

ECP_table_data <- ECP_scripts %>% 
  group_by(hb_name, drug) %>%  # Aggregate data by Health Board and drug
  summarise(
    total_quantity_4_years = sum(paid_quantity, na.rm = TRUE),  # Total prescriptions over 4 years
    avg_annual_total_quantity = total_quantity_4_years / 4,  # Average annual prescriptions
    hb_population = first(hb_population),  # hb_population is consistent within each hb_name
    .groups = "drop" # ungroup data
  ) %>%
  drop_na(drug) %>% # remove rows with no drug values
  mutate(avg_annual_presc_100000 = avg_annual_total_quantity * 100000 / hb_population) %>% #annual rate of prescriptions per 100,000
  select(hb_name, drug, avg_annual_presc_100000, hb_population) %>% # select columns before pivot
  pivot_wider(
    names_from = drug, 
    values_from = avg_annual_presc_100000, 
    values_fill = 0,  # Fill missing values with 0
    names_glue = "{drug}_rate"  # Rename columns for clarity
  ) %>% 
  clean_names() # clean drug_rate column names following pivot

# calculate total prescription rate per health board
ECP_table_data <- ECP_table_data %>% 
  mutate(total_ECP_rate = rowSums(select(., levonorgestrel_rate, ulipristal_acetate_rate), na.rm = TRUE)) %>%  # calculate total prescription rate for both drugs
  arrange(desc(total_ECP_rate)) # Arrange by total rate in descending order

#calculate percentage young people per health board to see if there is a trend between rate of prescribing of ECP and proportion of young people living in the HB
ECP_table_data_buckets <- ECP_table_data %>% 
  full_join(age_population_data) %>% 
  group_by(hb_name) %>% 
  mutate(prop_young_ppl_hb = pop_hb_16_34/hb_population) %>% 
  ungroup()

annual_avg_ECP_table <- ECP_table_data_buckets %>% 
  select(hb_name, levonorgestrel_rate, ulipristal_acetate_rate, total_ECP_rate,prop_young_ppl_hb) %>% # Select relevant columns
  gt() %>% 
  cols_label(hb_name = "Health Board",
             total_ECP_rate= "Total",
             levonorgestrel_rate= " Levonorgestrel",
             ulipristal_acetate_rate=" Ulipristal Acetate",
             prop_young_ppl_hb = "Percentage of young people") %>% # Rename the column
  # Format the columns with numbers to have two decimal places
  fmt_number(columns = c(levonorgestrel_rate, ulipristal_acetate_rate, total_ECP_rate, prop_young_ppl_hb), decimals = 2) %>% 
  
  # Centre the number columns using American spelling
  cols_align(align = "center",
             columns = c(levonorgestrel_rate,ulipristal_acetate_rate, prop_young_ppl_hb)) %>% 
  # Add an overall summary of the number columns
  grand_summary_rows(columns = c(levonorgestrel_rate,ulipristal_acetate_rate, total_ECP_rate),
                     fns = list("Overall Average" = ~mean(., na.rm = TRUE)),
                     fmt = list(~ fmt_number(., decimals = 2))) %>% 
  fmt_percent(columns = prop_young_ppl_hb,
              decimals = 2) %>%  # make the proportion of young people a percentage  
  # Add a title and subtitle; md() allows text formating from mark down 
  tab_header(title = md("**Average Annual Rate of Emergency Contraception Prescriptions by Health Board in Scotland**"),
             subtitle = md("This table presents the average annual prescription rates of emergency contraception pills (ECP) per 100,000 women, derived from the mean prescription rates across the years 2019 to 2022. Health Boards are ranked in descending order, starting with those with the highest prescription rates.")) %>%
  # use tab_spanner to add a title to each drug and total column.
  tab_spanner(
    label = md("*Prescription rate per 100,000 women*"),
    columns = c(levonorgestrel_rate,ulipristal_acetate_rate, total_ECP_rate)) %>% 
  tab_source_note(md("*Data from Public Health Scotland. Available from: (https://www.opendata.nhs.scot/dataset/prescriptions-in-the-community)*")) %>% 
  tab_stubhead(md("**2019-2022**")) %>% 
  tab_footnote(
    footnote = "includes Capital City, Edinburgh",
    locations = cells_body(columns = hb_name, rows = 2)
  ) %>% 
  tab_footnote(
    footnote = md("aged 16 to 34"),
    locations = cells_column_labels(columns = prop_young_ppl_hb)) %>% 
  opt_stylize(style = 6, color = "cyan")

annual_avg_ECP_table

Average Annual Rate of Emergency Contraception Prescriptions by Health Board in Scotland
This table presents the average annual prescription rates of emergency contraception pills (ECP) per 100,000 women, derived from the mean prescription rates across the years 2019 to 2022. Health Boards are ranked in descending order, starting with those with the highest prescription rates.
2019-2022	Health Board	Prescription rate per 100,000 women			Percentage of young people1
		Levonorgestrel	Ulipristal Acetate	Total
	NHS Greater Glasgow and Clyde	3,375.85	28.52	3,404.37	26.50%
	NHS Lothian2	2,494.97	56.37	2,551.34	27.32%
	NHS Ayrshire and Arran	2,027.24	15.29	2,042.53	18.98%
	NHS Tayside	1,364.97	22.08	1,387.05	22.29%
	NHS Forth Valley	1,268.46	73.31	1,341.78	21.94%
	NHS Grampian	1,210.23	15.88	1,226.10	22.44%
	NHS Fife	1,049.15	53.58	1,102.73	21.68%
	NHS Lanarkshire	783.56	47.97	831.54	21.52%
	NHS Shetland	778.66	17.45	796.11	19.11%
	NHS Orkney	609.26	6.72	615.98	17.47%
	NHS Dumfries and Galloway	514.87	24.25	539.12	17.42%
	NHS Western Isles	191.58	197.27	388.85	15.95%
	NHS Highland	284.97	52.14	337.11	17.46%
	NHS Borders	242.26	76.18	318.44	16.94%
Overall Average	—	1,156.86	49.07	1,205.93	—
Data from Public Health Scotland. Available from: (https://www.opendata.nhs.scot/dataset/prescriptions-in-the-community)
1 aged 16 to 34
2 includes Capital City, Edinburgh

How does deprivation on the type of emergency contraceptive prescribed

How does deprivation influence type of contraception prescribed? I wanted to explore if there was any correlation between the type of contraception being prescribed and deprivation. To do this I made a ratio of levonogestrel to total emergency contraception prescriptions.

A/(A+B) Where A = Levonogestrel (prescribed within 3 days of unprotected sex) B = Ulipristal Acetate (prescribed within 5 days of unprotected sex)

This is important to identify any variation in prescribing patterns in more deprived areas. Differences may suggest patients access health services later, so have to use Ulipristal Acetate, or that GPs or Pharmacists have differences in prescribing preferences in more deprived areas.

To measure deprivation I have used the Scottish Index of Multiple Deprivation. I took the SIMD 2020v2 dataset from Public Health Scotland website [available from: https://www.opendata.nhs.scot/gl/dataset/scottish-index-of-multiple-deprivation/resource/acade396-8430-4b34-895a-b3e757fa346e ] as this dataset contains SIMD decile weighted per Health Board population.

When interpreting SIMD in deciles rank 1 is considered the most deprived, and rank 10 is least deprived.

# read in datasets:

SIMD <- read_csv("https://www.opendata.nhs.scot/gl/dataset/78d41fa9-1a62-4f7b-9edb-3e8522a93378/resource/acade396-8430-4b34-895a-b3e757fa346e/download/simd2020v2_22062020.csv") %>%
  clean_names() %>%
  select(data_zone, simd2020v2hb_decile)

# I chose to use GP Practices and List sizes from October 2022, as this was the closest dataset I could find which correlated with the final year of my prescriptions datset
gp_addresses <- read_csv("https://www.opendata.nhs.scot/dataset/f23655c3-6e23-4103-a511-a80d998adb90/resource/1a15cb34-fcf9-4d3f-ad63-1ba3e675fbe2/download/practice_contactdetails_oct2022-open-data.csv") %>%
  clean_names() %>%
  select(practice_code, gp_practice_name, data_zone)

# Create ECP_GP by using the GP_addresses dataset to map the GP practice code to a datazone. I then used the column datazone to full_join() the SIMD dataset to the prescriptions dataset.
   
ECP_GP <- ECP_scripts %>%
  filter(!gp_practice %in% c(99996, 99997, 99998)) %>% # remove dummy GP practice codes as they do not have a known gp practice code so cannot be mapped to a SIMD.
  left_join(gp_addresses, by = c("gp_practice" = "practice_code")) %>%
  #left_join(data_zones, by = "data_zone") %>%
  left_join(SIMD, by = "data_zone") %>%
  drop_na(simd2020v2hb_decile) %>% # need SIMD value to make barchart
  group_by(gp_practice) %>% 
  mutate(total_quantity_per_gp = sum(paid_quantity)) %>% 
  clean_names()

# calculate the number of GPs per health board 
toal_no_GP_per_hb <- ECP_GP %>% 
    filter(!gp_practice %in% c(99996, 99997, 99998)) %>% # remove dummy GP practice codes as they do not have a known gp practice code so cannot be mapped to a SIMD.
  group_by(hb_name) %>%
  summarise(unique_gp_count_per_hb = n_distinct(gp_practice))
  
# calculate average population per GP in a health board 
pop_per_gp_hb <- population_data %>% 
  full_join(toal_no_GP_per_hb) %>% 
  mutate(avg_pop_per_GP = hb_population /unique_gp_count_per_hb)

# join pop_per_gp_hb to ECP_GP 
ECP_pop_per_gp_hb <- ECP_GP %>% 
  left_join(pop_per_gp_hb)

# make bar chart
ECP_SIMD_barchart <- ECP_pop_per_gp_hb %>%
  group_by(simd2020v2hb_decile, drug) %>%
  summarise(
    prescriptions_hb_gp = (total_quantity_per_gp *10000 / avg_pop_per_GP), # summarise the prescriptions per 100,000
    .groups = "drop"  # Ungroup data after summarisation
  ) %>%
  ggplot(aes(x = prescriptions_hb_gp, y = factor(simd2020v2hb_decile, levels = 1:10), fill = drug)) +
  geom_col() +
  scale_fill_brewer(palette = "Set2", name = "Drug Type") +  # add colour palette
  labs(
    title = "Barchart to show how the prescription rate of \n Emergency Contraceptive Pill varies by Scottish Index of Multiple Deprivation",
    subtitle = "This chart shows the average prescription rate of ECP per SIMD, \n accounting for differences in number of GPs in each SIMD.",
    x = "Prescriptions per 10,000 women",
    y = "SIMD Decile \n (1 = Most Deprived, 10 = Least Deprived)",
    fill = "Drug") +
  theme_minimal(base_size = 12) +
  theme(
    plot.title = element_text(size = 14, face = "bold", hjust = 0.5),
    plot.subtitle = element_text(size = 12, hjust = 0.5),
    axis.title.x = element_text(size = 12),
    axis.title.y = element_text(size = 12),
    axis.text.x = element_text(size = 10, angle = 45, hjust = 1),
    axis.text.y = element_text(size = 10),
    legend.position = "right",
    legend.title = element_text(size = 12, face = "bold"),
    legend.text = element_text(size = 10))

ECP_SIMD_barchart <- ggplotly(ECP_SIMD_barchart)
ECP_SIMD_barchart

Prescribing choices vary per regions

Data sets to consider exploring:

Conclusion

Recommendations from analysis

Limitations of the dataset and suggestions for future analysis

References

I used https://gsverhoeven.github.io/post/zotero-rmarkdown-csl/ to set up citations. Chosen BMJ style as common.